Absolute  Measure  of  Surface-conductivity  for  Heat.       393 
A  copper  ball,  2  centimetres  radius,  having  a  thermo-electric  junc- 
tion at  its  centre,  was  suspended  in  the  interior  of  a  double-walled 
tin-plate  vessel  which  had  the  space  between  the  double  sides  filled 
with  water  at  the  atmospheric  temperature,  and  the  interior  coated 
with  lamp-black.  The  other  junction  was  in  metallic  contact  with 
the  outside  of  the  vessel,  and  the  circuit  wras  completed  through  the 
coil  of  a  mirror  galvanometer.  One  junction  was  thus  kept  at  a  nearly 
constant  temperature  of  about  14°  Cent.,  while  the  other  had  the  gra- 
dually diminishing  temperature  of  the  ball. 
Having  adjusted  the  galvanometer  to  the  degree  of  sensitiveness 
desired,  the  copper  ball  was  heated  in  the  flame  of  a  spirit-lamp  till 
its  temperature  was  considerably  above  that  required  to  throw  the 
spot  of  light  off  the  scale;  it  was  then  put  into  position  in  the  inte- 
rior of  the  tin-plate  vessel,  and  as  soon  as  the  spot  of  light  came 
within  range,  the  deflections  from  the  zero  position  were  noted  at 
intervals  of  one  minute  exactly  till  the  change  of  deflection  was 
reduced  to  about  two  scale-divisions  per  minute. 
Two  series  of  experiments  were  made  in  this  way,  each,  consist- 
ing of  several  sets  of  readings.  In  the  first  the  ball  had  a  bright 
surface,  and  in  the  second  it  was  coated  with  a  thin  covering  of  soot 
from  the  flame  of  a  lamp,  and  in  both  the  air  was  kept  moist  by  a 
saucer  containing  a  quantity  of  water  placed  in  the  interior  of  the 
tin-plate  vessel. 
As  the  range  of  differences  of  temperatures  of  the  junctions  ex- 
tended over  50c  Cent.,  the  change  in  the  difference  of  thermo-electric 
qualities  of  the  copper  and  iron  wires  forming  the  junctions  was 
very  considerable,  and  it  was  necessary  to  make  a  careful  thermo- 
metric  comparison  of  the  temperatures  of  the  junctions  and  galvano- 
meter deflections.  For  this  purpose  the  junctions  were  tied  to  the 
bulbs  of  two  previously  compared  thermometers,  having  their  stems 
divided  to  tenths  of  a  degree  Cent. ;  these  were  then  placed  in  two 
vessels  of  water,  one  at  the  temperature  of  the  air,  and  the  other 
heated  by  small  additions  of  hot  water,  and  kept  well  stirred  ;  simul- 
taneous readings  of  the  thermometers  and  galvanometer  deflections 
were  then  taken  at  various  points  of  the  scale*,  from  which  the 
formula 
2/=0°-0924  +  0°-00C0227# 
substituting  sulphuric  acid  for  the  water  in  the  first ;  and  the  results  in  the  two 
cases  were  so  nearly  alike,  that  any  effect  due  to  the  moisture  or  dryness  of  the  air 
could  not  be  distinguished  from  errors  of  observation.  From  this  circumstance, 
as  well  as  the  limited  range  of  temperatures,  these  results  are  not  given  here. 
*  These  readings  were  plotted,  and  the  curve  drawn  through  the  points  agreed 
very  closely  with  a  portion  of  a  parabolic  curve  whose  equation  is 
y=2-4+10-6:r--019*2, 
y  denoting  the  deflections  of  the  galvanometer,  and  x  the  difference  of  tempera- 
1 0-fi 
ture;  y  is  a  maximum  when  x  =  —rr  =279°,  and,  the  colder  junction  having 
been  at  16°  Cent.,  we  get  295°  as  the  neutral  point  of  the  specimens  of  copper 
and  iron  wires  used — a  very  close  agreement  with  former  observations,  consider- 
ing the  great  distance  of  the  neutral  point  from  the  temperature  of  the  observations. 
